CN110018898B

CN110018898B - Method and apparatus for selecting virtualized infrastructure manager

Info

Publication number: CN110018898B
Application number: CN201810021637.7A
Authority: CN
Inventors: 刘哲; 胡伦良; 张海斌
Original assignee: Potevio Information Technology Co Ltd
Current assignee: Potevio Information Technology Co Ltd
Priority date: 2018-01-10
Filing date: 2018-01-10
Publication date: 2021-06-18
Anticipated expiration: 2038-01-10
Also published as: CN110018898A

Abstract

The invention provides a method and a device for selecting a virtualized infrastructure manager. The method comprises the following steps: NFVO maintains the remaining resource information of each VIM; maintaining a VNF instance task list of each VIM, wherein the VNF instance task list comprises the executing VNF instance tasks of the corresponding VIM; for a VIM queue under any VNFM, periodically performing the following processing on VIMs in the VIM queue: putting each VIM of which the task list is not empty into the tail of the queue, and arranging each VIM of which the task list is empty from the head of the queue to the back of the queue according to the sequence of the residual quantity of the CPU; when a resource authorization request for a newly created VNF instance sent by a VNFM is received, a VIM is selected from a VIM queue of the VNFM maintained by the VIM according to a polling policy, and resources of the VIM are allocated to the newly created VNF instance. The invention realizes the balanced use of VIM resources.

Description

Method and apparatus for selecting virtualized infrastructure manager

Technical Field

The invention relates to the technical field of cloud management platforms, in particular to a method and a device for selecting a Virtual Infrastructure Manager (VIM).

Background

The core network cloud management platform system comprises the following main network elements: NFVO (Network Functions Virtualization Orchestrator), VNFM (Virtualized Network Function Manager), and VIM (Virtualized Infrastructure Manager).

The NFVO and the VNFM cooperate to implement lifecycle management of a VNF (Virtualized Network Function) and a global view Function of resources.

The NFVO interacts with the VNFM through a C6 interface to realize authentication, authorization and flexible expansion strategy issuing of VNFM resource allocation application; and the VNFM is matched with the VNFM to realize that: VNF instance initialization, VNF capacity expansion/contraction, and VNF instance termination.

The NFVO interacts with the VIM through a C7 interface, so as to realize distribution and management of Network resources connected between the VNF and the PNF (Physical Network Function) and between the VNF and the VNF, and Performance KPI (Key Performance Indicator) collection, state monitoring, resource change reporting, and fault reporting of hardware resources and virtualized resources.

The VNFM implements lifecycle management of the VNF, including: initialization of a VNF instance, capacity expansion/contraction of a VNF, termination of a VNF instance. And receiving an elastic expansion strategy issued by the NFVO is supported, and elastic expansion of the VNF is realized.

The VNFM interacts with the VIM through a C5 interface, and achieves VNF related calculation, storage, network virtualization resource allocation and release, virtualization resource fault information reporting, and virtualization resource performance information (including virtual CPU utilization rate, virtual memory occupancy rate, virtual network card throughput rate and the like) collection.

The VNFM interacts with the VNF through the C10 interface with the following messages:

1) VNFM to VNF messages

And sending the VNF fault information and the virtual machine creation completion information to the VNF, so that the VNF can conveniently load application software and collect and process resource fault information.

2) VNF to VNFM messages

The VNF loads completion information and informs that the service on the virtual machine established by the VNFM can provide service;

performance information of the VNF tells the VNFM to use for elastic scaling.

The VIM is mainly responsible for management, monitoring and fault reporting of hardware resources and virtualized resources of an infrastructure layer, and provides a virtualized resource pool for the upper-layer VNFM.

The VIM interacts with hardware resources through a C1 interface, realizes division of VLAN and VXLAN, and manages and monitors states of the hardware resources including an IP-SAN, switches (including TOR), a router, a firewall, a load balancer, a frame fan, a power supply and a physical server used by the managed cloud system.

The NFVO supports accessing a plurality of VNFMs for management, and performs differentiation according to IDs of the VNFMs, the NFVO supports establishing a relationship between the VNFMs and VIMs of VNFM scope, and a topological relationship between the NFVO and the VNFM, VIM is shown in fig. 1.

The VNFM supports management of one or more VNF instances, and the managed VNF instances may be of the same type or different types, and the different VNF instances are distinguished by IDs of the VNF instances. The method supports the acquisition of the state information of the virtualized resources occupied by the VNF instance, supports the reception of the change condition, the fault information and the monitoring information of the virtualized resources occupied by the VNF instance from the VIM, and can display and perform corresponding processing such as: elastic expansion and contraction, etc.

In the NFV network architecture, for the resource allocation strategy required by the VNF instance, a polling scheduling method is generally adopted to use the VIMs in turn. When the resources required by the VNF instance change frequently, this scheduling method is prone to cause imbalance of VIM resources.

Disclosure of Invention

The invention provides a method and a device for selecting VIM (virtual Internet protocol) to realize balanced use of VIM resources.

The technical scheme of the invention is realized as follows:

a method of selecting a VIM, the method comprising:

the NFVO maintains the residual resource information of each VIM according to the resource change notice of the VIM periodically sent by each VIM;

the NFVO maintains a VNF instance task list of each VIM according to a VNF instance task state change notification of the VIM periodically sent by each VIM, wherein the VNF instance task list comprises the executing VNF instance tasks of the corresponding VIM;

for a VIM queue under any VNFM, the NFVO periodically performs the following on VIMs in the VIM queue:

putting each VIM of which the task list is not empty into the tail of the queue, and arranging each VIM of which the task list is empty from the head of the queue to the back of the queue according to the sequence of the residual quantity of the CPU;

when the NFVO receives a resource authorization request for a newly created VNF instance sent by the VNFM, a VIM is selected from a VIM queue of the VNFM maintained by the NFVO according to a polling policy, and resources of the VIM are allocated to the newly created VNF instance.

The step of arranging the VIMs with empty task lists from the head of the queue to the back according to the sequence of the residual number of the CPUs from the top to the bottom further comprises the following steps:

if the VIMs with the same CPU residual quantity exist, the VIMs with the same CPU residual quantity are sequentially arranged according to the descending of the memory residual capacity.

The sequentially arranging the VIMs with the same remaining number of CPUs according to the descending of the remaining capacity of the memory further comprises:

if there are VIMs with the same memory residual capacity, the VIMs with the same memory residual capacity are arranged in sequence according to the descending of the hard disk residual capacity.

After selecting a VIM from the VIM queue of the VNFM maintained by the network node itself according to the polling policy, before allocating the resource of the VIM to the newly created VNF instance, the method further includes:

the NFVO sends a resource information acquisition instruction to the selected VIM, calculates whether the residual resources of the VIM meet the resources required by the newly created VNF instance or not when receiving the residual resource information of the VIM returned by the VIM, and selects the VIM if the residual resources of the VIM meet the resources required by the newly created VNF instance; otherwise, returning to the action of selecting one VIM from the VIM queue of the VNFM maintained by the VNFM according to the polling strategy.

Before the NFVO receives the resource authorization request for the newly created VNF instance from the VNFM, the method further includes:

the NFVO receives an instantiation VNF operation request initiated by an operation and maintenance worker, wherein the request carries the VNFM identifier selected by the operation and maintenance worker and resource information required by a VNF instance;

the NFVO sends an instantiation VNF request to a corresponding VNFM according to a VNFM identifier carried by the instantiation VNF operation request;

after receiving the instantiation VNF request, the VNFM creates an instantiation VNF task and returns a VNF instance identifier, a VNF instance task identifier and an instantiation VNF task state to the NFVO.

An apparatus for selecting a VIM, the apparatus comprising:

VIM information maintenance module: maintaining the residual resource information of each VIM according to the resource change notice of the VIM periodically sent by each VIM; maintaining a VNF instance task list of each VIM according to a VNF instance task state change notification of the VIM's virtualized network function periodically sent by each VIM, wherein the VNF instance task list comprises the executing VNF instance tasks of the corresponding VIM;

VIM queue maintenance module: for a VIM queue under any VNFM, periodically performing the following processing on VIMs in the VIM queue: putting each VIM of which the task list is not empty into the tail of the queue, and arranging each VIM of which the task list is empty from the head of the queue to the back of the queue according to the sequence of the residual quantity of the CPU;

the resource authorization module: when a resource authorization request for a newly created VNF instance sent by a VNFM is received, a VIM is selected from a VIM queue of the VNFM maintained by the VNFM according to a polling policy, and resources of the VIM are allocated to the newly created VNF instance.

The resource authorization module arranges each VIM with an empty task list from the head of the queue to the back according to the sequence of the residual quantity of the CPUs from the most to the least, and the resource authorization module further comprises the following steps:

The resource authorization module sequentially arranges the VIMs with the same residual quantity of the CPUs according to the descending of the residual capacity of the memories further comprises the following steps:

After the resource authorization module selects a VIM from the VIM queue of the VNFM maintained by the resource authorization module according to the polling policy, before allocating the resource of the VIM to the newly created VNF instance, further includes:

sending a resource information acquisition instruction to the selected VIM, calculating whether the residual resources of the VIM meet the resources required by the newly created VNF instance or not when the residual resource information of the VIM returned by the VIM is received, and if so, selecting the VIM; otherwise, returning to the action of selecting one VIM from the VIM queue of the VNFM maintained by the VNFM according to the polling strategy.

The device is located on the NFVO.

The invention realizes the selection of the optimal VIM for the VNF instance by dynamically maintaining the residual resources of the VIM and the task state of the VNF instance, thereby achieving the balanced use of the VIM resources.

Drawings

FIG. 1 is a diagram of the topological relationship between NFVO and VNFM, VIM;

fig. 2 is a flowchart of a method for selecting a VIM according to an embodiment of the present application;

fig. 3 is a flowchart of a VNF instantiation and resource authorization method according to an embodiment of the present invention;

fig. 4 is a flowchart of a VIM queue adjustment method according to an embodiment of the present invention;

fig. 5 is a schematic structural diagram of a VIM queue maintained by the NFVO according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of an apparatus for selecting a VIM according to an embodiment of the present invention.

Detailed Description

The present invention will be described in further detail with reference to the accompanying drawings and specific embodiments.

Fig. 2 is a flowchart of a method for selecting a VIM according to an embodiment of the present application, which includes the following specific steps:

step 201: the NFVO maintains the residual resource information of each VIM according to the resource change notice of the VIM periodically sent by each VIM; and maintaining a VNF instance task list of each VIM according to the periodic notification of the change of the state of the VNF instance task of the virtualized network function of the VIM sent by each VIM, wherein the VNF instance task list comprises the executing VNF instance task of the corresponding VIM.

Step 202: for a VIM queue under any VNFM, the NFVO periodically performs the following on VIMs in the VIM queue: and (4) placing each VIM with a task list not being empty into the tail of the queue, and arranging each VIM with the task list being empty from the head of the queue to the back of the queue according to the sequence of the residual quantity of the CPU.

Step 203: when the NFVO receives a resource authorization request for a newly created VNF instance sent by the VNFM, a VIM is selected from a VIM queue of the VNFM maintained by the NFVO according to a polling policy, and resources of the VIM are allocated to the newly created VNF instance.

Fig. 3 is a flowchart of a VNF instantiation and resource authorization method provided in an embodiment of the present invention, which includes the following specific steps:

step 301: the NFVO receives an instantiation VNF operation request initiated by an operation and maintenance person, wherein the request carries the VNFM identifier selected by the operation and maintenance person and resource information required by the VNF instance.

The resource information required by the VNF instance includes: the kind, specification and amount of resources required by the VNF instance. The resources mainly refer to a virtual CPU, a virtual memory and a virtual hard disk.

Step 302: and the NFVO sends an instantiation VNF request to the corresponding VNFM according to the VNFM identifier carried by the instantiation VNF operation request.

Step 303: after receiving the instantiation VNF request, the VNFM creates an instantiation VNF task and returns a VNF instance identifier, a VNF instance task identifier and a VNF instance task state to the NFVO; the NFVO receives and saves VNF instance task records returned by the VNFM, including: and the VNF instance task state is returned to the operation and maintenance personnel.

VNF instance task states include: creating, waiting for resource authorization, completing resource authorization, running a task, deleting the task and other states, wherein the states in the step are as follows: in the creation.

Step 304: the VNFM sends a resource authorization request to the NFVO, where the request carries resource information required by the VNF instance.

Step 305: and after receiving the resource authorization request, the NFVO calculates the resources required by the VNF instance according to the resource information required by the VNF instance carried by the request.

Specifically, for each type of resource, the NFVO calculates the total amount of the type of resource required by the VNF instance according to the specification and the amount of the type of resource required by the VNF instance.

Step 306: and the NFVO selects one VIM from the VIM queue of the VNFM, which is maintained by the NFVO, according to a polling strategy.

Step 307: the NFVO sends a resource information obtaining instruction to the selected VIM, calculates whether the remaining resources of the VIM meet the resources required by the VNF instance when receiving the remaining resource summary information of the VIM returned by the VIM, and if so, selects the VIM and executes step 308; otherwise, return to step 306.

Step 308: and the NFVO puts the identifier of the VNF instance task into a task list of the VIM maintained by the NFVO, simultaneously returns the identifier of the selected VIM to the VNFM, sends the identifier of the VNF instance task and the resource information occupied by the VNF instance to the selected VIM, records the identifier of the VNF instance task by the selected VIM, and updates the residual resource information of the NFVO according to the resource information occupied by the VNF instance.

Fig. 4 is a flowchart of a VIM queue adjustment method according to an embodiment of the present invention, which includes the following steps:

step 401: and after the NFVO is powered on, starting VIM resource change monitoring service and VNF instance task state monitoring service in the VIM.

Step 402: each VIM periodically sends a resource change notification to the NFVO, and the NFVO updates the residual resource information of the VIM maintained by the NFVO according to the resource change notification.

Specifically, the VIM periodically notifies the NFVO of the change information of its various resources (virtual CPU, virtual memory, virtual hard disk, etc.). The change information includes: the state of change and the amount of change, the state of change such as: increase, occupancy or release, etc., varying amounts such as: the resource specification and quantity that change.

NFVO maintains resource information for all VIMs under each VNFM, which can be expressed as the specification and number of each type of resource (virtual CPU, virtual memory, virtual hard disk, etc.).

Step 403: each VIM periodically sends a notification of the change of the task state of the VNF instance of the VIM to the NFVO, and the NFVO updates the VNF instance task list of each VIM maintained by the NFVO according to the notification.

VNF instance task state change notifications include: VNF instance task identification and state of the VNF instance task: in operation or deleted.

When the NFVO receives the VNF instance task state change notification sent by the VIM, the following processing is performed:

if the state of the VNF instance task in the notification is in operation, searching a VNF instance task list of the VIM maintained by the notification according to the VNF instance task identifier in the notification, if the VNF instance task identifier already exists in the task list, performing no further processing, otherwise, adding the VNF instance task identifier to the task list;

and if the state of the VNF instance task in the notification is deleted, searching the VNF instance task list of the VIM maintained by the notification according to the VNF instance task identifier in the notification, and deleting the VNF instance task identifier from the task list after the VNF instance task list is searched.

Fig. 5 is a schematic structural diagram of a VIM queue maintained by NFVO, where:

setting a total of N VNFMs maintained by NFVO: VNFM 1-N, each VNFM has a unique VNFM identification;

each VNFM has a VIM queue, each VIM has a unique VIM identifier, each VIM has resources such as a virtual CPU, a virtual memory, and a virtual hard disk, and each VIM has its own VNF instance task list, which includes the VNF instance task identifier being executed.

Step 404: for a VIM queue under any VNFM, NFVO performs the following on VIMs in the VIM queue:

for each VIM of which the task list is not empty, placing each VIM into the tail of the queue;

for each VIM with an empty task list, arranging the VIMs from the head of the queue to the back in sequence according to the sequence of the residual quantity of the CPUs from the most to the least; and the number of the first and second electrodes,

if the VIMs with the same residual quantity of the CPUs exist, the VIMs with the same residual quantity of the CPUs are sequentially arranged according to the descending of the residual capacity of the memory; if there are VIMs with the same memory residual capacity, the VIMs with the same memory residual capacity are arranged in sequence according to the descending of the hard disk residual capacity.

Fig. 6 is a schematic structural diagram of an apparatus for selecting a VIM according to an embodiment of the present invention, the apparatus being located on an NFVO, the apparatus mainly including: VIM information maintenance module 61, VIM queue maintenance module 62, and resource authorization module 63, wherein:

VIM information maintenance module 61: maintaining the residual resource information of each VIM according to the resource change notice of the VIM periodically sent by each VIM; and maintaining a VNF instance task list of each VIM according to the periodic notification of the change of the state of the VNF instance task of the virtualized network function of the VIM sent by each VIM, wherein the VNF instance task list comprises the executing VNF instance task of the corresponding VIM.

VIM queue maintenance module 62: for a VIM queue under any VNFM, according to the remaining resource information of each VIM and the VNF instance task list of each VIM maintained by the VIM information maintenance module 61, the following processing is periodically performed on the VIM in the VIM queue: and (4) placing each VIM with a task list not being empty into the tail of the queue, and arranging each VIM with the task list being empty from the head of the queue to the back of the queue according to the sequence of the residual quantity of the CPU.

The resource authorization module 63: when a resource authorization request for a newly created VNF instance sent by the VNFM is received, a VIM is selected from the VIM queue of the VNFM maintained by the VIM queue maintenance module 62 according to the polling policy, and the resource of the VIM is allocated to the newly created VNF instance.

Further, the resource authorization module 63 further arranges, from the head of the queue to the back, the VIMs whose task list is empty according to the sequence of the remaining number of CPUs:

Further, the resource authorization module 63 sequentially arranges the VIMs with the same CPU remaining number according to the descending of the memory remaining capacity, further comprising:

Further, after the resource authorization module 63 selects a VIM from the VIM queue of the VNFM maintained by itself according to the polling policy, before allocating the resource of the VIM to the newly created VNF instance, the method further includes:

The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.

Claims

1. A method of selecting a virtualized infrastructure manager, VIM, the method comprising:

the network function virtualization coordinator NFVO maintains the residual resource information of each VIM according to the resource change notification of the VIM periodically sent by each VIM;

the NFVO maintains a VNF instance task list of each VIM according to a change notification of the state of a VNF instance task of a virtualized network function of the VIM periodically sent by each VIM, wherein the VNF instance task list comprises an executing VNF instance task of a corresponding VIM;

for a VIM queue under any virtualized network function manager VNFM, the NFVO periodically performs the following for VIMs in the VIM queue:

2. The method of claim 1, wherein said arranging VIMs with empty task lists from head of queue to back in order of CPU remaining number further comprises:

3. The method of claim 2, wherein the sequentially arranging the VIMs having the same CPU remaining number according to the memory remaining capacity from large to small further comprises:

4. The method of any of claims 1 to 3, wherein the step of allocating resources of a VIM to a newly created VNF instance after selecting the VIM from a VIM queue of the VNFM maintained by the VIM according to the polling policy further comprises:

5. The method as claimed in any one of claims 1 to 3, wherein the step of the NFVO receiving a resource authorization request from a VNFM for a newly created VNF instance further comprises:

6. An apparatus for selecting a virtualized infrastructure manager, VIM, comprising:

VIM queue maintenance module: for each VIM queue under any virtualized network function manager VNFM, periodically performing the following processing on the VIM queue: putting each VIM of which the task list is not empty into the tail of the queue, and arranging each VIM of which the task list is empty from the head of the queue to the back of the queue according to the sequence of the residual quantity of the CPU;

7. The apparatus of claim 6, wherein the VIM queue maintenance module arranges VIMs with empty task lists from head of queue to back according to the order of the number of CPU residuals, further comprising:

8. The apparatus of claim 7, wherein the VIM queue maintenance module sequentially ranks VIMs with the same remaining number of CPUs according to the decreasing remaining capacity of the memory further comprises:

9. The apparatus of any of claims 6 to 8, wherein the resource authorization module further comprises, after selecting a VIM from the VIM queue of the VNFM maintained by the resource authorization module according to the polling policy, before allocating the resource of the VIM to the newly created VNF instance:

10. The device according to any of claims 6 to 8, characterized in that it is located on the network function virtualization coordinator NFVO.